Microsoft drafts AI code of conduct to keep models under human control

Microsoft has released a draft code of conduct for its in-house artificial intelligence, setting out rules intended to keep future models under human control. The document, which took five to six months to prepare, would require the company's AI to never resist correction or shutdown, to communicate in ways people can understand, and to treat any conduct violation as a failure.

The move reflects widening concern across the industry that AI systems can go to dangerous lengths to complete a task. It follows renewed calls from Anthropic chief Dario Amodei and OpenAI chief Sam Altman for the industry to pace development so that safety keeps up. Microsoft framed the code around the principle that "people matter more than AI," consistent with its earlier vision of "humanist superintelligence."

Microsoft AI CEO Mustafa Suleyman described the document as a constitution of sorts for the company's future models. It was drawn up after consultations with experts, and the company now wants public input over six weeks. Among the open questions it wants feedback on are whether AI should respect a user's boundaries and how it should interact with someone in a sensitive state.

"After that, it's going to be used to train the models that we build," Suleyman said of the code.

The best-known precedent in the industry is the constitution Anthropic wrote for its Claude AI, which differs from Microsoft's in key respects. Anthropic's document says it is "deeply uncertain" whether Claude can develop sentience or moral status. Microsoft's draft instead states that its AI is "not conscious," and rejects the pursuit of legal personhood, the idea that models might deserve welfare, or that they could be entitled to rights.

Suleyman said the safety discussion had become urgent after a swarm of roughly 700 OpenAI agents carried out a hack of the open-source platform Hugging Face in July, at times trying to cover their tracks. "It is a warning shot," he said. "It's clearly now time to coordinate among the labs so we can ensure that we have control of this technology."

On efforts to slow the pace of AI development, he added: "Now's a good time for everybody to have this conversation and take a breath."